From 31bcd36e4658c7a418797058ce5fcc57da0d87fc Mon Sep 17 00:00:00 2001 From: Rahul Somani Date: Tue, 14 Dec 2021 19:34:04 +0530 Subject: [PATCH 1/2] add tinynet models --- timm/models/efficientnet.py | 94 +++++++++++++++++++++++++++++++++++++ 1 file changed, 94 insertions(+) diff --git a/timm/models/efficientnet.py b/timm/models/efficientnet.py index 3d50b704..ec7e17c7 100644 --- a/timm/models/efficientnet.py +++ b/timm/models/efficientnet.py @@ -23,6 +23,10 @@ An implementation of EfficienNet that covers variety of related models with effi * Single-Path NAS Pixel1 - Single-Path NAS: Designing Hardware-Efficient ConvNets - https://arxiv.org/abs/1904.02877 +* TinyNet + - Model Rubik's Cube: Twisting Resolution, Depth and Width for TinyNets - https://arxiv.org/abs/2010.14819 + - Definitions & weights borrowed from https://github.com/huawei-noah/CV-Backbones/tree/master/tinynet_pytorch + * And likely more... The majority of the above models (EfficientNet*, MixNet, MnasNet) and original weights were made available @@ -407,6 +411,22 @@ default_cfgs = { url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_m-0f4d8805.pth'), 'tf_mixnet_l': _cfg( url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_l-6c92e0c8.pth'), + + "tinynet_a": _cfg( + input_size=(3, 192, 192), # int(224 * 0.86) + url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_a.pth'), + "tinynet_b": _cfg( + input_size=(3, 188, 188), # int(224 * 0.84) + url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_b.pth'), + "tinynet_c": _cfg( + input_size=(3, 184, 184), # int(224 * 0.825) + url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_c.pth'), + "tinynet_d": _cfg( + input_size=(3, 152, 152), # int(224 * 0.68) + url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_d.pth'), + "tinynet_e": _cfg( + input_size=(3, 106, 106), # int(224 * 0.475) + url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_e.pth'), } @@ -1140,6 +1160,50 @@ def _gen_mixnet_m(variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrai return model +def _gen_tinynet( + variant, model_width=1.0, depth_multiplier=1.0, pretrained=False, **kwargs +): + """Creates a TinyNet model. + """ + arch_def = [ + ['ds_r1_k3_s1_e1_c16_se0.25'], ['ir_r2_k3_s2_e6_c24_se0.25'], + ['ir_r2_k5_s2_e6_c40_se0.25'], ['ir_r3_k3_s2_e6_c80_se0.25'], + ['ir_r3_k5_s1_e6_c112_se0.25'], ['ir_r4_k5_s2_e6_c192_se0.25'], + ['ir_r1_k3_s1_e6_c320_se0.25'], + ] + model_kwargs = dict( + block_args = decode_arch_def(arch_def, depth_multiplier, depth_trunc='round'), + num_features = max(1280, round_channels(1280, model_width, 8, None)), + stem_size = 32, + fix_stem = True, + round_chs_fn=partial(round_channels, multiplier=model_width), + act_layer = resolve_act_layer(kwargs, 'swish'), + norm_layer=kwargs.pop('norm_layer', None) or partial(nn.BatchNorm2d, **resolve_bn_args(kwargs)), + **kwargs, + ) + + features_only = False + model_cls = EfficientNet + kwargs_filter = None + + if kwargs.pop('features_only', False): + features_only = True + # kwargs_filter = ('num_classes', 'num_features', 'head_conv', 'global_pool') + kwargs_filter = ('num_classes', 'num_features', 'conv_head', 'global_pool') + model_cls = EfficientNetFeatures + + model = build_model_with_cfg( + model_cls, variant, pretrained, + default_cfg=default_cfgs[variant], + pretrained_strict=not features_only, + kwargs_filter=kwargs_filter, + **model_kwargs) + if features_only: + model.default_cfg = default_cfg_for_features(model.default_cfg) + + return model + + @register_model def mnasnet_050(pretrained=False, **kwargs): """ MNASNet B1, depth multiplier of 0.5. """ @@ -2209,3 +2273,33 @@ def tf_mixnet_l(pretrained=False, **kwargs): model = _gen_mixnet_m( 'tf_mixnet_l', channel_multiplier=1.3, pretrained=pretrained, **kwargs) return model + + +@register_model +def tinynet_a(pretrained=False, **kwargs): + model = _gen_tinynet('tinynet_a', 1.0, 1.2, **kwargs) + return model + + +@register_model +def tinynet_b(pretrained=False, **kwargs): + model = _gen_tinynet('tinynet_b', 0.75, 1.1, pretrained, **kwargs) + return model + + +@register_model +def tinynet_c(pretrained=False, **kwargs): + model = _gen_tinynet('tinynet_c', 0.54, 0.85, pretrained, **kwargs) + return model + + +@register_model +def tinynet_d(pretrained=False, **kwargs): + model = _gen_tinynet('tinynet_d', 0.54, 0.695, pretrained, **kwargs) + return model + + +@register_model +def tinynet_e(pretrained=False, **kwargs): + model = _gen_tinynet('tinynet_e', 0.51, 0.6, pretrained, **kwargs) + return model From 450ac6a0f5bfd2db4d30355c99a23b91d9ab63e4 Mon Sep 17 00:00:00 2001 From: Ross Wightman Date: Tue, 21 Dec 2021 23:51:54 -0800 Subject: [PATCH 2/2] Post merge tinynet fixes for pool_size, feature extraction --- timm/models/efficientnet.py | 51 ++++++++++++------------------------- 1 file changed, 16 insertions(+), 35 deletions(-) diff --git a/timm/models/efficientnet.py b/timm/models/efficientnet.py index ec7e17c7..b1c570b2 100644 --- a/timm/models/efficientnet.py +++ b/timm/models/efficientnet.py @@ -413,19 +413,19 @@ default_cfgs = { url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_l-6c92e0c8.pth'), "tinynet_a": _cfg( - input_size=(3, 192, 192), # int(224 * 0.86) + input_size=(3, 192, 192), pool_size=(6, 6), # int(224 * 0.86) url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_a.pth'), "tinynet_b": _cfg( - input_size=(3, 188, 188), # int(224 * 0.84) + input_size=(3, 188, 188), pool_size=(6, 6), # int(224 * 0.84) url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_b.pth'), "tinynet_c": _cfg( - input_size=(3, 184, 184), # int(224 * 0.825) + input_size=(3, 184, 184), pool_size=(6, 6), # int(224 * 0.825) url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_c.pth'), "tinynet_d": _cfg( - input_size=(3, 152, 152), # int(224 * 0.68) + input_size=(3, 152, 152), pool_size=(5, 5), # int(224 * 0.68) url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_d.pth'), "tinynet_e": _cfg( - input_size=(3, 106, 106), # int(224 * 0.475) + input_size=(3, 106, 106), pool_size=(4, 4), # int(224 * 0.475) url='https://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_e.pth'), } @@ -1172,35 +1172,16 @@ def _gen_tinynet( ['ir_r1_k3_s1_e6_c320_se0.25'], ] model_kwargs = dict( - block_args = decode_arch_def(arch_def, depth_multiplier, depth_trunc='round'), - num_features = max(1280, round_channels(1280, model_width, 8, None)), - stem_size = 32, - fix_stem = True, + block_args=decode_arch_def(arch_def, depth_multiplier, depth_trunc='round'), + num_features=max(1280, round_channels(1280, model_width, 8, None)), + stem_size=32, + fix_stem=True, round_chs_fn=partial(round_channels, multiplier=model_width), - act_layer = resolve_act_layer(kwargs, 'swish'), + act_layer=resolve_act_layer(kwargs, 'swish'), norm_layer=kwargs.pop('norm_layer', None) or partial(nn.BatchNorm2d, **resolve_bn_args(kwargs)), **kwargs, ) - - features_only = False - model_cls = EfficientNet - kwargs_filter = None - - if kwargs.pop('features_only', False): - features_only = True - # kwargs_filter = ('num_classes', 'num_features', 'head_conv', 'global_pool') - kwargs_filter = ('num_classes', 'num_features', 'conv_head', 'global_pool') - model_cls = EfficientNetFeatures - - model = build_model_with_cfg( - model_cls, variant, pretrained, - default_cfg=default_cfgs[variant], - pretrained_strict=not features_only, - kwargs_filter=kwargs_filter, - **model_kwargs) - if features_only: - model.default_cfg = default_cfg_for_features(model.default_cfg) - + model = _create_effnet(variant, pretrained, **model_kwargs) return model @@ -2277,29 +2258,29 @@ def tf_mixnet_l(pretrained=False, **kwargs): @register_model def tinynet_a(pretrained=False, **kwargs): - model = _gen_tinynet('tinynet_a', 1.0, 1.2, **kwargs) + model = _gen_tinynet('tinynet_a', 1.0, 1.2, pretrained=pretrained, **kwargs) return model @register_model def tinynet_b(pretrained=False, **kwargs): - model = _gen_tinynet('tinynet_b', 0.75, 1.1, pretrained, **kwargs) + model = _gen_tinynet('tinynet_b', 0.75, 1.1, pretrained=pretrained, **kwargs) return model @register_model def tinynet_c(pretrained=False, **kwargs): - model = _gen_tinynet('tinynet_c', 0.54, 0.85, pretrained, **kwargs) + model = _gen_tinynet('tinynet_c', 0.54, 0.85, pretrained=pretrained, **kwargs) return model @register_model def tinynet_d(pretrained=False, **kwargs): - model = _gen_tinynet('tinynet_d', 0.54, 0.695, pretrained, **kwargs) + model = _gen_tinynet('tinynet_d', 0.54, 0.695, pretrained=pretrained, **kwargs) return model @register_model def tinynet_e(pretrained=False, **kwargs): - model = _gen_tinynet('tinynet_e', 0.51, 0.6, pretrained, **kwargs) + model = _gen_tinynet('tinynet_e', 0.51, 0.6, pretrained=pretrained, **kwargs) return model